Microbes in our guts influence our metabolism, moods, and behaviors, but the problem of understanding how these influences arise is demonstrably complex. 100 trillion cells from 1000 species with millions of genes make up the human microbiome. Just as the one gene = one function paradigm has largely evaporated from the field of genetics in favor of understanding how pathways of interactions lead to a phenotype, the field of microbiology has largely begun to recognize that ecology is at the core of many microbiome disease states. Ecology means that a web of biotic and abiotic factors interact to produce a system-level output. One of the core concepts that has developed the field of ecology is the 'keystone species'. In a food web (network map of the interactions between species), keystone species interact with many more species than the average species does and these species have reverberating effects on an ecosystem when they are eliminated, such that the stability of the ecosystem often fails and many other species are eliminated by indirect effects due to loss of the keystone species. One of the toughest problems in treating ailments of the microbiome is that microbiomes themselves are robust to change. While antibiotics can kill off the vast majority of microbes, when the flora recover, they usually represent the same flora the patient started with. The only widely successful change of the microbial ecosystem in patients is through the use of fecal transplants, whereby the entire gut flora of a patient is replaced with a donor's stool using an enema. My aim is to use the keystone species concept as a strategy by which to perturb the gut flora without eliminating them entirely. By mapping the microbial food web, I aim to determine candidate keystone species. By developing targeted bacteriophage therapies against the keystone candidates, I aim to restructure microbial food webs to change the metabolic output, thus affecting the core metabolites that affect host metabolism, mood, and behavior. I will approach the project from two angles: (i) I will establish a model, reduced complexity gut microbiome in the fruit fly, which is ideal for studying behavioral outputs (ii) I will examine full- complexity gut microbiomes in humanized mouse guts through a collaboration with a gnotobiotic mouse facility to test fundamental principles established in the fly system from a more human- relevant perspective.